There are four types of redundancy that one can use on ASAs. The first one you cited, redundant interfaces on a single physical device is the least common in my experience.
The second is failover - when the ASA is mated is a failover ASA in a high availability configuration. This is the most common usage for customers requiring high availability (HA). That is the most common implementation and has been around since ASA 7.0 software (i.e. a good many years).
The third is to bond your interfaces from a given ASA (or sets of interfaces if you have an HA pair) into an Etherchannel. This has the added advantage of giving you potentially higher trhoughput. Etherchannel support was introduced in ASA software version 8.4(1).
The fourth and newest method is clustering. It was introduced just last fall in ASA 9.0 and is not very widely adopted just yet. It is primarily for high throughput requirements exceeding a single device's capacity but also gives the added benefit of redundancy.
None of them require you setup things twice configuration-wise. Some file operations (software upgrade, certificate management, VPN profiles (XML files)) need to be copied onto both members in a failover pair or all members in a cluster scenario.
Edit - there is a fifth type specific to VPNs whereby one can configure a secondary VPN gateway for clients, usually at a alternate site. That approach does require settting up everything separately on the ASAs.
